The Playground is a sandbox for one-time conversations with answer engines in Genezio. These test conversations do not change your visibility metrics or your recommendation metrics. With the Playground, you can examine, test, and show demos. The measurements that the rest of the platform uses stay clean.

​
Why the Playground keeps tests separate from measurement

Genezio runs scenarios on a daily schedule. Each of these scenarios gives data to your AI Recommendations, your AI Visibility, and the other KPIs of the platform. This is the correct behavior for measurement. But it is not correct when you want to:
  • Test how a new persona behaves before you use it in a scenario
  • Do a one-time investigation (“let me just ask ChatGPT this one thing as a 35-year-old marketer in Germany”)
  • Try a different text before you make it a real scenario
  • Show a live demo without a change to the numbers of the customer
The Playground keeps tests separate from measurement. The runs in the Playground do not go into your reported metrics.

​
What you can configure in a Playground run

When you start a run in the Playground, you select:
  • The answer engine (each supported answer engine)
  • The persona of the run (with its language and its location)
  • The agent type: Prompter, Recommender, Comparer, or Introspector. The agent type controls the structure of the conversation.
  • The language of the run
Then, Genezio runs a full multi-step agent conversation, as it does for a real scenario. The only difference is that Genezio does not collect metrics from it.

​
Where to find the Playground

You can open the Playground from the Settings screen. You can also load a specific scenario into the Playground from the scenario detail view. Do this to run an existing scenario again with a different persona, answer engine, or text. The metrics of the initial scenario do not change.

​
Playground history and automatic removal

Genezio saves each run of the Playground in a history view. Thus, you can:
  • Compare runs side by side
  • Go back to an old run
  • Change the persona, the text, or the answer engine, and try again
Genezio automatically removes old messages of the Playground to keep the history clean. The Playground is a work sandbox, not a long-term archive. When a run gives a result that you want to keep, do one of these actions:
  • Make it a real scenario.
  • Write the results in a brief.
  • Send it to the Geo Assistant for analysis.

​
Typical Playground workflows

Test a new persona before you use it. Write a draft persona. Run some Playground conversations with it, and see how answer engines respond. When the results are correct, make it a real persona. Examine a hypothesis. For example: “What does ChatGPT recommend in our category to a buyer with constraint X?” Start one Playground run and read the answer. If necessary, change the question and try again. Test a scenario change before you save it. Load an existing scenario into the Playground and change the text. Make sure that the response of the answer engine is as you expect. Then, save the change to the real scenario. Show a demo. During a sales demo or a demo for stakeholders, show live responses of answer engines for the brand of the customer. The measured numbers of the customer do not change.

​
Playground compared with scheduled scenarios